Prediction Modeling for Tropical Signalgrass Emergence in Florida

Prediction Modeling for Tropical Signalgrass (Urochloa subquadripara) Emergence in Florida

Travis C. Teuton, Christopher L. Main, Thomas C. Mueller, Department of Plant Sciences, University of Tennessee, Knoxville 37996; John B. Wilkerson, Department of Biosystems Engineering and Environmental Science, University of Tennessee, Knoxville 37996; Barry J. Brecke, Department of Agronomy, University of Florida, Milton 32583; and J. Bryan Unruh, Department of Horticultural Sciences, University of Florida, Milton 32583 37996

Corresponding author: Travis C. Teuton. tteuton@utk.edu

Introduction

Tropical signalgrass is a dominant weed in the St. Augustinegrass industry of Florida and is potentially troublesome from movement with St. Augustine sod and mechanical seed dissemination. Tropical signalgrass tolerates most selective postemergence herbicides labeled for application to St. Augustinegrass (4). Preemergence herbicide options such as oryzalin or benefin + oryzalin control tropical signalgrass for 8 to 11 weeks after treatment (4).

In Florida, the majority of tropical signalgrass can be found south of U.S. Interstate Highway 4 (latitude 28°N). Tropical signalgrass usually does not perenniate north of US Interstate Highway 4 due to killing freezes (4); however, annual re-infestation from seed is common in these areas. In our previous research, tropical signalgrass emergence was observed when weekly mean soil or ambient air temperatures were ≥ 20°C (3).

Geographic Information Systems (GIS) modeling of tropical signalgrass emergence is possible since tropical signalgrass germination parameters are known and historical weather data (air temperature) is available from the National Oceanographic and Atmospheric Administration (NOAA) (1). A GIS prediction model based on historical weather data, similar to the crabgrass prediction model by Main et al. (2), would benefit producers to ensure optimum preemergence herbicide timing for control of tropical signalgrass. Our objective was to develop a tropical signalgrass emergence map based on historical weather data to predict the spring emergence of tropical signalgrass.

Building the Prediction Map

Methods used for GIS interpolation were similar to those proposed by Main et al. (2). Historical weather data was obtained from NOAA Daily Station Normals (1971-2000) for 103 weather stations across Florida (Fig. 1) (1). Temperature data and station location were entered into a database. An interpolated surface for tropical signalgrass emergence was developed for Florida using ESRI Arc Map 8.2 (ESRI, Redlands, CA) software with the spatial analysis extension. Weekly mean temperatures were calculated beginning on the first day of January. Kriging interpolation method (4) was used to predict weekly mean air temperature values for entire state using 103 sample points. A semi-variorum was used to determine the range, partial sill, and nugget values (parameters required for this GIS method) for weekly temperature means. Each week was subsequently interpolated by Ordinary Kriging (default settings) and a temperature layer for each week from February to May was created for Florida.

Fig. 1. Predicted emergence pattern of tropical signalgrass throughout Florida by Ordinary Kriging, including weather station locations and the Ft. Lonesome research location.

Since the environmental requirement for tropical signalgrass germination is a weekly mean soil or air temperature of 20°C (3), layers were analyzed to determine the first week-long period that had an average high temperature of ≥ 20°C. The first week meeting this criteria (week of January 1-7) was set as the starting date for further analysis. After Kriging interpolation, no differences were observed in temperature prediction from the first week of analysis to the week of February 8-15 (second week of February), therefore the initial starting date for further analysis was set as the week of February 8-15.

Beginning with the February 8 layer, data were reclassified to distinguish between areas with average temperatures above and below 20°C. Each layer then represented discrete variables (18 total) that determined tropical signalgrass emergence. Discrete layers were combined to create a map and then smoothed. The map indicated areas that had weekly mean temperatures of 20°C beginning on February 8 thru April depicting the predicted pattern of tropical signalgrass emergence across Florida.

Interpolation of Weather Data

Temperatures in the coastal areas of Dade and Broward Counties remain above 20°C. Thus tropical signalgrass can emerge at any time during the year (Fig. 1). However, predicted emergence of tropical signalgrass in inland areas of Dade and Broward Counties does not begin until the second week of February. Conversely, tropical signalgrass emergence in temperate areas of northwest Florida is not predicted to occur until April 22 through 29.

Application timing of preemergence herbicides should be planned one to two weeks in advance of the ideal germination date since these herbicides are only active on germinating seedlings. Our earlier growing degree-day model recommended that preemergence herbicide applications be made during the first or second week of March in Ft. Lonesome (3). This GIS model predicted emergence at the Ft. Lonesome site would occur during the third week of March and thus a preemergence herbicide application in the second week of March would be ideally timed for management of tropical signalgrass.

These results demonstrate that Kriging interpolation of historical weather data and known tropical signalgrass seed germination parameters investigated in a field setting can be used to create emergence prediction maps. These prediction maps can be useful in developing an IPM plan to manage weeds. Applying GIS to biological data for tropical signalgrass emergence transforms data collected out of one Florida location (Ft. Lonesome) into a tropical signalgrass management plan for the entire state of Florida.

Literature Cited

1. National Oceanic and Atmospheric Administration. 2001. Daily station normals 1971-2000. NOAA, National Climatic Data Center, Asheville, NC.

2. Main, C. L., Robinson, D. K., McElroy, J. S., Mueller, T. C., and Wilkerson, J. B. 2004. A guide to predicting spatial distribution of weed emergence using geographic information systems (GIS). Online. Applied Turfgrass Science doi:10.1094/ATS-2004-1025-01-DG.

3. Teuton, T. C., Brecke, B. J., Unruh, J. B., MacDonald, G. E., Miller, G. L., and Tredaway Ducar, J. 2004. Factors affecting seed germination of tropical signalgrass (Urochloa subquadripara). Weed Sci. 52:376-381.

4. Teuton, T. C., Unruh, J. B., Brecke, B. J., MacDonald, G. E., Miller, G. L., and Tredaway Ducar, J. 2004. Tropical signalgrass (Urochloa subquadripara) control with preemergence and postemergence applied herbicides. Weed Technol. 18:419-425.